Recent Research at GFDL on Surface Temperature Trends and Simulations of Tropical Cyclone Activity in the Indian Ocean Region - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 53

start dates up into the late 1980s according to the results shown. When the
black region lies within the pink (or violet) shaded regions, the observed trend
is termed “consistent” with the All Forcing runs. This is the case for all start
dates shown. When the observed trend is both detectable and consistent with
the All Forcing (or if it is significantly greater than the All Forcing trend), we
term this a case of a detectable anthropogenic signal, where anthropogenic
forcing has contributed, at least in part, to the observed trend. This is the case
for trends with start dates up into the late 1980s. In short, a detectable
anthropogenic warming is clearly present in the Indian Ocean SST data,
according to our model assessment.
What is described for Fig. 2 above is a categorization of the observed
trends. This same categorization approach can be applied at the level of
individual model grid points. Specifically, for a given start year of a trend to
2010, the observed trend can be categorized as: (i) no detectable trend, (ii)
detectable warming that is less than simulated, (iii) detectable warming that is
consistent with All Forcing runs (i.e., attributable and consistent anthropogenic
warming), or (iv) detectable warming that is greater than in the All Forcing
runs (i.e., attributable anthropogenic warming that is greater than simulated).
Figure 3 shows the mapped results of such an analysis for the trends in annual
mean data over the period 1901-2010. The map indicates that the IPCC AR5
(CMIP5) historical All Forcing runs have a warming trend in the Indian Ocean
since 1900 that is similar to that observed. The categorization assessment
indicates that there is some tendency for the model ensemble to simulate too
small a warming rate since 1901 in the North Indian Ocean (dark red shading),
and too high a warming rate in parts of the South Indian Ocean within 30
degrees of the equator (yellow-orange shading). Note that there is at least a
detectable warming trend over almost the entire Indian Ocean since 1901,
according to this analysis. Much of this warming is either consistent with (red)
or exceeds (dark red) the warming in the All Forcing ensemble, and is
inconsistent with Natural Forcing only, indicating that a detectable
anthropogenic trend is very likely present in the data according to our analysis.
This anthropogenic warming signal also extends over the Indian subcontinent
and parts of south and southeast Asia.
Figure 4 shows the results of doing a similar trend categorization analysis
for individual three-month seasons (i.e., December-February; March-May; June-
August; and September-November). The under-prediction of the warming trend
in the All Forcing runs is most pronounced in the boreal summer (June-August)
and fall (September-November) seasons. Even with the reduced data associated
with doing a seasonal analysis, an anthropogenic warming trend contribution
is still detectable over most of the Indian Ocean region (red or dark red shading),
and over much of the Indian subcontinent, except for the March-May season
and to some extent June-August.